SOURCE: "Hafez and Poetic Unity through Verse Rhythms," Journal of Near Eastern Studies, Vol. 31, No. 1, January, 1972, pp. 1-10.

[In the following essay, Hillman attempts to explain the musical elements of Hafiz's verse, contending that it is the inability of translators to adequately capture these rhythms in English that makes their work unsatisfying.]

The following is an impressionistic translation of a poem by Hafez (fourteenth century A.D.), the premier lyric poet in the Persian language:

(1) Your entwined tresses,
I'm ever drunk with the brought breeze of
  them.
Your sorcerer's eyes,
I'm ever lost to self because of them.
(2) My long vigil—lord—oh,
Will it raise a night time vision
Of your brows, my vigil shrine,
Before which the candles of my eyes glow—
(3) Their black orbs I hold dear:
Mirrors of your midnight Hindu mole.
(4) The world is yours to immortalize:
Bid the breeze blow aside your veil.
(5) The world is yours to eternalize:
Pray the breeze rain down
Life down from among your hair.
(6) The breeze and I are one,
Homelessly blown about.
Drunk with your hair's fragrance
Is the breeze,
I with your sorcerer's eyes.
(7) Hafez's lofty love be lauded—
His lot here and hereafter
That his eye cherish threshold dust
And no more.²

This rendering of a typical Hafezian lyric poem³ departs further, in terms of conventional verse aspects of the original, from the Persian than do Edward Fitz-Gerald's quatrains from the Khayyamic robais from which FitzGerald drew inspiration; yet, it is closer in spirit and tone to Hafez's original than is Matthew Arnold's Sohrab and Rustum to Ferdowsi's Sohrab.⁴ However, the fact remains that FitzGerald and Arnold composed successful poems in English, thus—as has not been done with Hafez's poems—bringing to readers of English poetry a taste of Persian robai and literary epic traditions at their best.

That Hafez has not survived translation at the hands of the dozen or more scholars and lovers of medieval Persian poetry who have published versions of Hafez's poems in English may be less the result of the fact that no one as talented as FitzGerald and Arnold has tried his hand at it,⁵ than of the fact that there is inherent in the verse form employed by Hafez an emphasis on the musical potentiality of words, phrases, and verse lines that resists translation. In fact, in some of Hafez's poems, the patterns of verbal rhythm seem to be a basic organizing principle, the plain sense of the poetic statements complementing the verbal music.

A partial elucidation of this asserted emphasis or reliance by Hafez on the unifying power of musical elements of verse is the aim of this paper, the thrust being not to determine how, in terms of it, Hafez ought to be translated, but rather to demonstrate that the poetics which can be deduced from Hafezian poems such as Q95 seems quite different from the poetics of lyric poetry in the English language.⁶

To this end, the following literally translated version of Q95, freely rendered above in an attempt to convey something of the integrity and intensity of the speaker's feeling and tone, is presented. The literal version, while losing all the life of the original, at the same time preserves its images, metaphors, and plain sense with exactitude (words in parentheses are explanatory, indicating synonyms, connotations, or ordinary second denotations):

1. The breeze (fragrance) from (of) the lock(s, curls) of your hair intoxicates me continually (wine). The deceit of your magician's (sorcerer's) eye(s) destroys (annihilates, brings to loss of awareness of self) me continuously (time and time again, second by second; i.e. look by look).
2. After all this endurance (patience, perseverance), oh lord, will (can) one see (envision) an evening on which I light (my) candle-(like) eye(s) in the mehrab (prayer-niche or alcove in a mosque, i.e., the direction in which people pray, facing the mehrab) of your eyebrow(s) (to which the curved upper edge of the mehrab is similar in shape).
3. The blackness of the surface of my eye's pupil(s) I hold dear (preserve) because it is'a memento (written copy) for my soul of the picture (image) of your Indian mole (which is, thus, reflected in the lover's pupils which function as a mirror).
4. If you want to adorn the world wholly (completely) with immortality, tell the breeze to blow aside for a moment (time) the veil from your face;
5. and, if you wish to abrogate the law (destroy the custom) of impermanence (transience, mortality) from the world, shake (your head) till (so that) from each (strand of) hair thousands of spirits (souls, life) fall (pour) to the ground (down).
6. The breeze and I (are) two destitute (musk-bearing) and perplexed (wandering, aimless) (creatures), (our efforts) fruitless; I (am) intoxicated by the sorcery of your eye(s); and he (the breeze) is (drunk) with the fragrance of your hair.
7. Praise be (how wondrous is) the ambition Hafez has: in this world and in the next nothing will be cherished by his eye(s) except for the dirt (dust) at your threshold.⁷

The two translations taken together reveal, first, a consistency of tone. The speaker's attitude toward the beloved and toward his own plight vis-a-vis the beloved is one throughout the poem. Secondly, the images in the poem are of a piece. The lover's state is that of a worshiper, a petitioner, the beloved, by virtue of her perfection, effecting in him the conditions of intoxication, perhaps spiritual, and the loss of awareness of self, analogous to spiritual absorption. Thirdly, a single theme runs through the whole poem. The beloved is a subjectively ideal and unattainable beauty, in the service of whose love the speaker worshipfully and fruitlessly remains faithful. Whether or not the poem thematically oscillates between a subjective idealization of the love for a woman and the possible use of the image of a woman to represent the condition of man's love for his creator is a moot point, Hafez's possibly intentional ambivalence in this regard in no way affecting the integration of the tone, images, and theme, but merely offering another level upon which the poetic materials may be felt to operate.⁸

However, whatever can be said for the consistency of tone, theme, and imagery in this poem, nothing about Hafez's use of a particular tone, theme, and images persuades a reader that the poet's ordering of his statements (numbered in both translations above) is inevitable or final. For example, nothing evidenced by the literal translation above prevents a reader from rear-ranging the numbered statements as follows:

1, 6, 4, 5, 2, 3, 7
1, 6, 2, 3, 4, 5, 7
4, 5, 1, 6, 2, 3, 7
2, 3, 1, 6, 4, 5, 7

To be sure, these suggested reorderings indicate some limitations in the capacity of Hafez's poem to submit to rearrangement. For example, the seventh statement is, in terms of its content, a concluding note. The third statement continues the image and sense of the second; and, the coordinating conjunction at the beginning of the fifth statement links it to the fourth.…

This capacity of Hafez's poem, and many others in the collected poems of Hafez,⁹ to submit to a hypothetical rearrangement of its several statements without violation of the tone, theme, or patterning of images seems to set much medieval Persian lyric poetry apart, as a species, from lyric poetry in English and prompts many readers of poems such as Q95 to judge medieval Persian lyric poetry defective (i.e., in terms of so-called organic unity) in comparison with lyric poetry in English.¹⁰

In any case, it is obvious that if Hafez's Q95 is a poem, that is, is a single, whole poetic statement ordered through means proper to poetry, the principles of its organization must lie in the words and statements of the poem as seen other than as aspects of the plain sense, tone and theme.…

From this transliteration, it is seen that the poem contains fourteen verses with a rhyme scheme of a, a, b, a, c, a, d, a, e, a, f, a, g, a. The rhyme scheme itself reveals three facts. First, the second verse of every pair of verses exhibits a feminine rhyme in -uyat (-at = "your …," "of you"), the capacity for such a monorhyme in Persian where the rhyme strikes a hearer as neither forced nor monotonous quite unparalleled in English verse. And, if perchance an English poet succeeds once or twice in using such a monorhyme, then let the limit be raised to a hundred such pairs of verses, which some medieval Persian panegyrics contain, or let the rhyme be expanded to seven syllables (i.e., six repeated syllables following a syllable of masculine rhyme), as Hafez himself employs in Q75.

Secondly, in structural terms, the rhyme scheme bears out the previous observations on the endstopped nature of the poem's several statements, each statement now seen to be a pair of verses, and on the lack of inevitability or finality in the arrangement of these pairs of verses. For, if the traditional sonnet rhyme schemes abba, abba, cdecde and abab, cdcd, efef gg generally reflect the divisions or aspects of thoughts, feelings, images, and statements in that verse form, the rhyme scheme in Hafez's poem implies that each pair of verses is coordinate to the other pairs, there not being a more lengthy or comprehensive rhyme-thought-syntactic unit than one pair.

Thirdly, the first pair of verses consists of two rhymed verses (-suyat, -duyat), which gives this pair special aural emphasis. That being so, this pair of verses naturally deserves a place of special emphasis in the poem, either at the beginning or at the end, those being the usual two places of special emphasis in discourse. But, this rhymed pair of verses cannot occupy the final slot because the seventh pair of verses, as observed above, very obviously constitutes a concluding note to the poem, part of its force as such the effect of the use of the nom de plume, Hafez, in that pair. Therefore, the aa pair naturally occupies the first place in the poem, its internal rhyme a means of arousing reader interest and creating in an auditor an expectation of the repetition of the rhyme a in later verses.

If the above is a fair analysis, the pattern of the poem is seen finally to be:

    2, 3
1 + 4, 5 + 7
      6

the arrangement of 2-3, 4-5, and 6 with respect to each other still arbitrary. This deduced pattern: first pair (rhymed) + x-number of pairs + last pair (nom deplume) represents, in fact, the limits of any traditional definition of the Persian verse form called ghazal of which Hafez's Q95 is a typical example. For instance, in Alessandro Bausani's words, a Persian ghazal

… consists of a few bayts (verses, or distichs), generally not less than five and no more than twelve, with a single rhyme (often accompanied by a radif); in the first bayt, called matla, both hemistichs too [sic] rhyme together; the last bayt, called makta, contains the nom-de-plume (takhallus) of the author; the contents of the ghazal are descriptions of the emotions of the poet in front of [sic] love, spring, wine, God, etc., often inextricably connected.¹³

Thus, if a principle of organization exists in the Persian ghazals which, like Hafez's Q95, are not ordered on the basis of theme or its manner of presentation, images or their patterning, or the tone of the poetic speaker, the conventions of the ghazal verse form do not provide a clue either. On the other hand, there is no question that a critical reader of ghazals, such as Q95, senses in the original a singleness of impression and a completeness and integrity of poetic statement, in short, the same sense that he experiences in reading good lyric poems in English.

What accounts for this sense seems, very simply, to be the existence in such Persian ghazals of a set of rhythmic patterns so basic, pervasive, and subtly intertwined as to weld the non-rhythmic aspects of the bayts together and to help create the effects called poetic.¹⁴ Thus, in Q95 not only does it seem impossible to determine the arrangement of the three sets of couplets 2-3, 4-5, and 6, but it seems unnecessary as well insofar as the obligation of the groups of bayts is to make contextual sense and their function, beyond that, to provide the raw materials for the resultant patterns of rhythm.

Vocabulary in some Persian ghazals is seen to be manipulated as are musical notes and chords. The pervasiveness of end rhyme in medieval Persian poetry (deriving from) the anterior, causative facility for rhyme inherent in Persian) is one piece of evidence for this assertion. A stronger piece of evidence is the often observed fact¹⁵ that a poet such as Hafez employs a totally conventional poetic direction which flourished from four centuries before Hafez and that is employed even today by some poets. In other words, there is no vocabulary item in Q95 that does not appear, and in the same context, in the lyric poems of countless Persian poets before Hafez's time and countless times in his own poetry. For the reader of English poetry who does not know Persian any attempt at verbalization of this fact partakes of understatement because the capacity of a reader of English poetry to conceive of fresh and effective poetic statement involving a word, an image, a bayt, or a whole ghazal that five or six poets have previously made use of is experientially limited. But, the freshness and effectiveness of Q95, in which the word "breeze," for example, appears for about the two hundredth time in the course of Hafez's collected poems, one hundred fifty of these occasions in similar terms and contexts, is a fact. And, such poetic freshness, if it is often in Hafez a matter of "what oft was thought but ne'r so well expressed," nevertheless owes most to the verbal music that conventional vocabulary, images, metaphors, sequences of images, and conventional themes serve to create, this music creating the distinctiveness of individual poems which employ conventional vocabulary and figures of speech.…

Because the Persian language, like English, is characterized by patterns of syllable stress in words and phrases, and does not feature syllables long or short by nature or position, the adoption of the Arabic ilm alarud (science of prosody) by the early Persian poets of the Islamic era obliged these poets and later prosodists to formulate strictly Persian zahaf (deviations from a basic or "regular" metric foot) in using Arabic meters, to make greater use of some more suitable, "regular" meters infrequently employed in Arabic, and to conventionalize metrical fictions in the use of even the most basic and suitable Arabic meters.¹⁷ …

In any case, even such a fiction is regularly observed, the result being that, although the rhythmic quality of a poem such as Q95 in a standard Arabic meter is far different from the quality of an Arabic poem in the same meter, in which the quantity of a syllable (i.e., a long syllable takes twice as long to pronounce as a short syllable) takes precedence over word and rhetorical stress patterns, Q95 can be read in metronomic fashion according to the hazaj meter.¹⁹ And, unquestionably, the hazaj rhythm emerges in any natural reading of the poem, even if read by a person unaware of the existence of the Arabic meter employed. A reader incapable of scanning Q95 quantitatively might immediately sense any imperfection or deviation from the hazaj meter if such an imperfection existed.

But, the hazaj meter in Q95 Jies behind a stronger accentual pattern of rhytham …, according to which every verse exhibits five stressed syllables vis-à-vis differing numbers of unstressed syllables (each verse has fifteen to seventeen syllables). Other readings of Q95 are possible; but each would seem to fall into an accentual pentameter pattern. This pattern, of course, corresponds to the nature of the Persian language which, as mentioned above, is characterized by patterns of syllable stress in single words and phrases. What the most basic verse rhythm in Persian may be or what the most natural foot is composed of, that is, analogous to the consideration of the iamb in English verse, this writer does not know, there existing no discussions of Persian meters in terms of stress patterns or accents in Persian literature studies.²⁰ It is possible that a basic Persian foot, reflecting the most elemental accent patterns of multisyllabic Persian words and Persian phrases in converation does not exist as such, because of a difference in effect between the accentual rhythm that is at the core of medieval Persian poetry and that which a poet in English might compose in. The difference is that the regular number of stressed syllables in a Persian verse tends to drown out the unstressed syllables so that a hearer is not conscious of irregular amounts of space, in terms of the number of pronounced unstressed syllables, between each stressed syllable; whereas, the very term "sprung rhythm" in English seems to imply a sort of leaping out of a stressed syllable in the face of unstressed syllables, their relative number varying and recognized as being varied. In short, the basic aural effect of this sprung rhythm in Persian is a sense of strong lilt and regularity that would, in fact, border on sing-song, did not the less natural and native, yet better understood and appreciated, quantitative meters, such as the hazaj meter observable in Q95, exist, rising contrapuntally to the surface, especially in groups of unstressed syllables (some of which are bound to be conventionally long in quantity) that occur between stresses.

In conclusion, it would seem that Hafez's Q95, according to any formalistic criticism, is a composition in which three statements, i.e., 2-3, 4-5, and 6, are free to move about within the framework of the opening a a couplet and the concluding couplet featuring the nom de plume.²¹ The effects of the monorhyme, marked alliteration and assonance patterns, and euphony of the composition, which embodies a single theme, tone, and consistency of imagery, is to lend an external or superficial unity to the composition. But, it is the combination of the hazaj quantitative tetrameter superimposed on the sprung rhythm pentameter of syllable stress that effects the inner or organic integration and oneness and pleasure that is common to poetry. When T. S. Eliot observes that, "It is the immediate favorable impression of rhythm and diction which disposes us to accept a poem, encourages us to give it further attention and to discover other reasons for liking it,"²² he hits the mark, but from the opposite side of the target. For, his experience with Western poetry teaches him to look for something, usually meaning, beyond the music; whereas in Q95 it is the music beyond the meaning that is to be looked for. On a first reading, Q95 appeals to a reader because of its plain sense and images and promises him that there is more aurally to be discovered in further readings, in which the warp and woof of the patterns of rhythm gradually communicate themselves.

It is this writer's contention that Q95, and others like it in Hafez's Divan, is a poem, and a successful one, by virtue of its interwoven patterns of rhythm, an aspect of its character that wholly resists transmission in translation and that, perhaps, has resisted transmission in the foregoing description of it. In any case, the realization that very different means were at Hafez's disposal to achieve the same ends and effects achieved by lyric poets in English may prompt new critical approaches through which medicval Persian lyric poetry can be better understood and appreciated and through which a fuller sense of Persian originals can be transmitted by scholars who try their hand at producing English translations.

Notes

¹ Several ideas in this paper were first aired in a paper read at the First Iranology Congress of the University of Tehran (September 1970), the text of that paper subsequently printed in Ayandegan 3 (1970-71), No. 857 (October 15, 1970), p. 4, and No. 858 (October 18, 1970), pp. 4 and 7. I wish hereby to thank Professors Hosayn Nasr and Iraj Afshar for allowing me to address the "Iranian Literature—Islamic Period" Section of the Congress.

² Mohammad Qazvini and Qasem Ghani, eds., Divane … Hafez (Tehran: Zavvar, 1941), No. 95, p. 66. The variant recorded in this edition, based upon the oldest complete MS of Hafez's poems yet published in Iran (Khalkhali MS, dated 827 A.H.), corresponds exactly in number and order of verses with that recorded in an MS selection of Hafez's poems, dated 813 and 814 A.H., edited and published by Parviz Natel Khanlari, Ghazalha-ye … Hafez Shirazi (Tehran: Sokhan, 1958), p. 30. The significance of this correspondence of variants in the criticism of Hafez's poems, considered by M. C. Hillmann, "Naqd-e Adabi va Divan-e Hafez," Rahnema-ye Ketab 13 (1970-71): 44-45, lies in the fact that, because Hafez apparently did not prepare or approve an edition of his poems, only those poems the variant texts of which correspond in the oldest MSS in terms of number and order of verses are ready for critical scrutiny.

³ Mohammad Ali Eslami, "Taamol dar Hafez," Yaghma 16 (1963-64): 218, observes that, although "… it is one of Hafez's simpler ghazals, without any complicated philosophical or mystical subject matter, … it is one of those ghazals that can well serve as a demonstration of Hafez's thought and poetic craft." This essay, reprinted in Eslami's Jam-e Jahan-bin, 3d ed. (Tehran: Ebn-e Sina, 1971), pp. 261-80, has been translated by M. C. Hillmann in an article called "Sound and Sense in a Ghazal of Hafez," The Muslim World 61 (April 1971): 111-21.

⁴ An examination of wherein essentially Arnold's poem differs from Ferdowsi's is the subject of an article by M. C. Hillmann entitled "Rostam: A Study of Probability in Ferdowsi and Arnold," to appear in Iran-Shinasi 3 (1971), No. 1. In any case, Arnold never read any of Ferdowsi's Shahnameh, even in translation.

⁵ A. J. Arberry, "Hafiz and his English Translators," Islamic Culture 20 (1946): 111-28 and 229-49, entertains the merits and defects of extant English translations of Hafez, as does part of Arberry's "Orient Pearls at Random Strung," Bulletin of the School of Oriental and African Studies 11 (1946): 699-712. A list of Hafez translators is included in Arberry's compilation, Fifty Poems of Hafez (Cambridge: University Press, 1962), p. 62, to which list must be added Peter W. Avery and John Heath-Stubbs, Hafiz of Shiraz: Thirty Poems (London: John Murray, 1952) and Robert Rehder's versions in Anthology of Islamic Literature (New York: Mentor, 1966), edited by James Kritzeck.

⁶ As Elder Olson argues in "Sailing to Byzantium: Prolegomena to a Poetics of the Lyric," reprinted in Wilbur Scott, ed., Five Approaches of Literary Criticism (New York: Collier Books, 1962), pp. 216-17, the question is properly one of the poetics of individual lyric poems, their "scrutiny" being "the beginning of the critical enterprise" which inquires "what form has been imposed upon the medium of words." Such inductive openmindedness is especially necessary in comparative literature study where the greater familiarity of a critic with his native poetic tradition may tend to lead him unwittingly to employ the limits of that tradition as a criterion in the study of other poetic traditions.

⁷ This interpretation differs slightly, for textual reasons (cf. note 12 below), from those by Eslami, "Taamol" pp. 275 ff. and by the Turkish Sudi (sixteenth century A.D.), Hafez's most famous commentator, in Sharh-e Sudi bar Hafez, vol. 1, translated by Esmat Sattarzadeh (Tehran: Rangin, 1963), No. 100, pp. 583-87.

⁸ According to Mahmud Human, with Esmail Khui, Hafez, 2d ed. (Tehran: Tahuri, 1968), pp. 110, 117-18, and 351-52, Q95 is an example of the first period of Hafez's development as a poet and thinker, a period in which the poet, in Human's view, is primarily concerned with secular love, rendi ("nonconformist abandon"), and nazarbazi ("ocular flirtation").

⁹ Within the spectrum of Hafez's collected poems, viewed in terms of formal or organic unity, Q95 seems to lie about half-way between a poem such as Q26 (translated in Arberry, Fifty Poems, p. 90f., which submits to no transposition of its several statements, and the Western favorite, Q3 (translated in Arberry, ibid., pp. 85-86), which gives voice to three or four distinct themes in its nine statements. On the basis of a consideration of poems such as Q3, Arberry, in "Orient Pearls," p. 704 f., asserts that Hafez "constructs each lyrie poem upon the basis of a limited number of themes selected from a repertory which is itself definitely restricted, and to a great extent conventional," the arrangement of the selected themes giving the key to an understanding of the poem's form. In the case of a poem such as Q95, which embodies a single tone, one theme, and consistent imagery, yet submits to a hypothetical transposition of its several statements, Arberry's explanation does not apply and, perhaps, a fortiori, does not apply to poems embodying several distinct themes.

¹⁰ Ali Akbar Dehkhoda, "Ghazal," Loghat Nameh, No. 62 (Tehran: University Press, 1961), p. 209, asserts that the thematic independence of each bayt (pair of verses) in a ghazal (the verse form exemplified by Hafez's lyric poems) is a technical flaw inherent in the verse form. Walter Leaf, as quoted in Arberry, "Orient Pearls," p. 704, remarks: "In the Persian ode (ghazal) we find a succession of couplets often startling in their independence, in their giddy transitions from grave to gay, from thought to mood. To the Persian each couplet is a whole in itself, a nokteh or point, sufficiently beautiful if it be adequately expressed, and not of necessity owing anything or adding anything to that which comes before or after. It is from the common metre and common rhyme alone that the ode gains a formal unity…" Leafs insinuations are (1) that rhyme and meter are practices with the same purposes and effects regardless of the nature of the language in the poetry of which they may be employed; (2) that a "common metre and common rhyme alone" are not enough to effect real formal unity; and (3) that, owing to the supposed independence of individual couplets in some medieval Persian lyric poems, the whole poem can be assumed to have only that effect which is the sum of the several effects of the couplets read individually.…

¹³ Alessandro Bausani, "Ghazal," Encyclopedia of Islam: New Edition (London: Luzac, 1965), p. 1033, in which essay Bausani asserts that a conceptual incongruity among the statements in a single poem seems to be a formal characteristic of the technical ghazal form: this, however, begs the question.

¹⁴ What these effects are seems unamenable to definition, although a reader of poetry recognizes them experientially. As T. S. Eliot, On Poetry and Poets, p. 6, puts it, ". poetry has to give pleasure. If you ask me what kind of pleasure then I can only answer, the kind of pleasure that poetry gives."

¹⁵ Shebli Nomani, Sher al-Ajam, vol. 2, 2d ed., translated by Mohammad Taqi Fakhr Gilani (Tehran: Ebne Sina, 1950), pp. 204 and 211; E. G. Browne, A Literary History of Persia, vol. 2 (Cambridge: University Press, 1956), p. 84.

¹⁶ Gotthold Weil, "Arud," Encyclopedia of Islam, New Edition, pp. 667-77, is the most comprehensive description of Arabic prosody available in English. The available English translations of medieval studies of Arabic prosody as adapted by Persian prosodists and poets include: H. Blochmann, The Prosody of the Persians (Calcutta: Lewis, 1872) and G. S. Ranking, The Elements of Arabic and Persian Prosody (Bombay: Education Society's Press, 1885), both books containing a translation of Sayfi's "Essay on Prosody," Blochmann's study including, as well, a translation of Jami's "Essay on Rhyme."

¹⁷ The existence of some zahaf is, as Mahdi Hamidi, in Aruz-e Hamidi (Tehran: Haydari, 1963), pp. 7-27, sardonically exhibits, no more than the admission that the metrical facts of classical Persian verse are different from or more comprehensive than the conventional Arabic molds which are supposed to account for all the possibilities. In other words, the conventional classification of meters in ilm al-arud does not represent some possible and natural Persian metrical patterns (aside from the fact that these meters do not reflect the nature of Persian verse as essentially accentual). Q95, however, was chosen for discussion partly because it exhibits salem (i.e., sound or regular) hazaj feet. P. N. Khanlari describes the inapplicability of Arabic arud to classical Persian verse in greater detail in the second chapter of Vazn-e Sher-e Farsi (Tehran: Bonyade Farhang-e Iran, 1966), pp. 93-108. Khanlari's study is the first comprehensive attempt to derive a valid theoretical explanation of what are the bases and kinds of classical Persian meters. Unfortunately, the attempt is marred by linguistic inaccuracies and imprecision, as, for example, the author's failure to consider the differences between word, rhetorical, and metrical accents. Also, in his consideration of the quantitative meters that may operate in classical Persian verse (i.e., not the theoretically applicable meters of Arabic origin, but the ones actually employed), Khanlari arbitrarily chooses all the mathematical possibilities, save one, of combining so-called long and short syllables in dimeter and trimeter feet (p. 161) to represent the basic patterns, rather than to conduct an inductive analysis of poetic practice.…

¹⁹ Meredith-Owens, loc. cit., observes: "The most outstanding feature of the Arud system as adopted by the Persians is the emphasis laid on quantity, which gives to Persian verse a lilt and swing which can be more readily appreciated by ears to which the more subtle rhythms of Arabic verse are unfamiliar." What Meredith-Owens fails to recognize is -that "the lilt and swing" probably owes nothing to the adoption of Arabic meters by Persian poets; but, it is rather the natural, accentual pattern of the Persian language, described below in relation to Q95, emerging as the basic rhythm upon which the Arabic quantitative meters are superimposed. Also, the subtlety of Arabic verse rhythms unquestioned, that of Persian, insofar as subtlety is related to complexity, ought to be tentatively established by its description in this essay.

²⁰ Both Hamidi, Aruz-e, and Khanlari, Vazn-e Sher, loc. cit., by demonstrating the inadequacies of the Arabic arud system as a description of medieval Persian poetic practice, bring the investigation of medieval Persian prosody to the point where the question of stress patterns should naturally present itself. However, to this writer's knowledge, the question has not heretofore been discussed, despite the fact that it may be the single, most important critical problem in the study of medieval Persian technical ghazals insofar as no feature or aspect of many ghazals can account for the reader's sense that they are poems except for the patterning of rhythms.

²¹ Of course, the statements are not "free to move about" in the sense that an editor may determine their order according to his own taste, because, whatever the reason, conscious or intuitive, Hafez ordered Q95 in a particular way. William K. Sumner rightly asks whether musical movement beyond the two verse patterns might not make the order of 2-3, 4-5, and 6 even hypothetically unchangeable. However there does not seem to be any musical movement beyond individual verses and couplets, although, in line with T. S. Eliot's observation, in his On Poetry and Poets, p. 30, that "the music of verses is not a line by line matter but a question of the whole poem," the rhythms in Q95 have a developing, cumulative effect.

²² Ibid., p. 191.